Alternative to LED encoder rings: illuminated red-green encoders?

[latigid on]

Just a quick concept with not too much thought behind it:

 

LED rings are great but seem very clunky to implement. You either need a single board (à la Fairlightii) with through hole LEDs or surface mount, which introduces panel problems (i.e. one needs a transparent case). Or you design stackable daughter boards like FantomXR, a good idea but adds to the cost.

 

An alternative might be to light the encoders directly. For example, Sparkfun have these encoders with a switch button and two common cathode LEDs (e.g. red+green). Currently $2.66-$2.95 depending on quantity.

 

https://www.sparkfun.com/products/10596

 

 

 

 

I have to admit I don't have a lot to go on from the software side, but I imagine one could get some pretty intuitive colour mixing by changing the levels of each red and green LED. If we consider that there are usually 16 LEDs per ring, a possible 2 bit scheme could look like this.

 

Green Red
0     0       OFF

1     0       green colours
2     0
3     0

3     1       orange colours
3     2

1     1       
2     1
3     3
1     2
1     1

1     2 
1     3

0     1       red colours
0     2
0     3

 

Of course it might be swapped around somewhat, but it's a general idea to start.

 

Interested to know how this might work from the software side because I think for bling factor and hardware implementation it could be quite worthwhile. The colours also enable other possibilities, like a VU meter scheme, LFO phase, simple on/off switch or 4-state parameters etc.

 

 

 

 

[FantomXR]

Or you design stackable daughter boards like FantomXR, a good idea but adds to the cost.

 

It's all about the quantity. I don't think that this is very expensive if I order 1.000 pc of pcb assembled LED-ring-PCBs. But what should I do with 1.000 pc? :D

 

Anyway: Your concept sounds great.

[novski]

Hi

I made my 8Enc PCB arround that Encoder.

Its quite interresting. Then a year later i desided to make a Matrix modul and a encoder mounting pcb. Its not yet on Github because i was to busy and still waiting for the big sister with ledring to arrive but you are welcome to use it...

The parts laks documentation as well because of time but im working on it...

Have a look: http://www.vlrlab.com/home/29-bourns-rgb-encoder-pcb.html

Im mot able to ship cheep because everything in switzerland is so expensive, as well as the parts.. I buy them for almost double the price that sparkfun sells them from the supplyers in switzerland...

If you have any question regarding them please fell free to contact me or have a look into the schematics of 8enc on Github.

Best Regards, novski

[FantomXR]

Hey Novski,

 

I checked your website again. I saw that you use LED-Rings as well. So for what purpose do you use the LEDs which are integrated in the encoder? 

[latigid on]

Good to hear that the encoders are nice! Indeed, the interface is there, but what is the implementation with a DOUT module?

 

Looks like Duggle has experimented with PWM on DOUT:

 

 

More recent one:

Edited July 12, 2015 by latigid on

[novski]

It wil help to visualize the Bank. I didn't realize it until now because i got stuck with the Motorfader modul. I try to make a remote for my software mixer. So in one bank it will be the Gain and in a other maybe the panorama and so on, i think the color will make it more visual in what mode i am while turning the encoder...

For a more complet feature request, the solution cold be to assign RGB LED's to a Ring (i think It is not possible yet to assign a look of a Matrix in RGB way...) and then give it to the RGB LED of the encoder.

Edited July 12, 2015 by novski

[Phatline]

great idea... that would make a lot of easier....maybe not on the programming side... i think as soon you learn how to read that colours it works great.

 

as expirience the alps caps are a bit big... so the saved space between the encs is away, but have no alternative in moment.

[latigid on]

great idea... that would make a lot of easier....maybe not on the programming side... i think as soon you learn how to read that colours it works great.

 

as expirience the alps caps are a bit big... so the saved space between the encs is away, but have no alternative in moment.

 

 

For the programming I suppose there are benefits in avoiding matrix routing (also easier for PCB routing, just use separate DOUT pins for each LED). And for the space, couldn't we use Waldorf-style clear knobs? I imagine some room could be gained over the LRE8x2, even by staggering the encoders (i.e. offsetting a row horizontally).

 

BTW, if someone would like to do the programming, I'm keen to do the PCB, maybe for a "Wilba-style" all-in-one MIDIbox CV v2. It might include 16 encoders with switches, CLCD, 4xGLCD and a few extra buttons. If not, maybe this would be a good place for me to start :).

[novski]

Yes, my tests resulted in a distance of ~30mm to the next encoder to get enough space for fingers and cap. My 8Enc pcb has a 27,5mm grid. With 11 LEDs in the Ring and the RGB colored Encoder.

[latigid on]

Depends on how many encoders one wants to cram on a panel :) Is 16 enough? (Fairlightiii used 43 mm spacing on a 342 mm PCB)

[FantomXR]

My grid on the encoder PCB is 29mm. This is because I want to add OLEDs underneath it and they have a width of 27,3mm. My encoder pcb can be stacked with a SMD LED-Ring (0805) which has 16 LEDs. 

 

Maybe I'll add the support for those RGB-Encoders. 

Edited July 13, 2015 by FantomXR

[latigid on]

Hmm, there's a few examples of LED driver ICs too:

 

MAX7221 is $10 oer chip...

 

TLC5940 is $3-6 but PDIP marked as end of life.

 

Also Ander's crazy Station, using PCA9635 ($2.50 in TSSOP):

[latigid on]

Getting closer:

 

[latigid on]

Just thinking about this a bit more, RGB encoders are known (and by the way: supported by MIOS DOUT to 7 levels + off, so no I2C buss/chips needed), but perhaps a bit too expensive. To make matters worse, it looks like a lot of them are common anode *facepalm*.

 

Another idea is to find a way of making a "collar" that sits over the shaft with the knob on top. On the PCB side a few RGB LEDs would illuminate this "lightpipe". So it could work out a bit cheaper and more flexible this way.

 

See this "Star Knob", a bit big, but the concept is about right.

 

 

 

http://www.okw.co.uk/en/Tuning-knobs/Star-Knobs.htm

[latigid on]

Another option is finding hollow shaft encoders and sticking a normal through hole RGB LED in.

[rhdf]

The idea having a color of the cap as indication of where you are seems a bit.. odd.
For personal use. Ok. But how to describe for others where, say 50% is?

For led-rings the now discontinued  http://mayhewlabs.com ledring, using a tlc5925 would be a great start.
With some redesign you can  get a boardlayout that is not even near the " bulkyness" of the original.

I My self started to layout a board with a couple of encoders and surrounding led-rings. Guess what takes the most of the PCB-space? The leds of course. So the obvious soulution would be a slimmed led-ring PCB mounted to the "main board"

 

 

 

Edited August 29, 2015 by rhdf

[latigid on]

I would say that we don't know how it will work until somebody (i.e. me) tries it. You have 256 colour levels with RGB LEDs but only 16 with the LREs, although you do have a positional indicator. Remember that exact values can be read out on the associated LCD. So hopefully I have some time to design and test after the BLM is out the door.

Interested to see what you come up with, both in terms of electronics (see FantomXR's work above) and also the "engineering" side (panel spacing, getting the light through etc.).

[novski]

Hi

If your working with Eagle i have a lib where i tested different angles and sizes...
Hope it helps..

Regards, novski

Bourns Encoder.lbr

[latigid on]

Heh, I see this is more about the LED ring but it's still interesting to see the RGB encoder...

Edited August 30, 2015 by latigid on

[rhdf]

Interested to see what you come up with, both in terms of electronics (see FantomXR's work above) and also the "engineering" side (panel spacing, getting the light through etc.).

Well, the electronics bit will probably be   more or less the same as the mayhew-labs ledring. 

My idea is to have a ~270 degree arc, aprox. 5mm wide cut out in the panel, and then insert a piece of cnc-cut plexi. that will serve as "lightpipe" and diffusor (if wanted)

Hi

If your working with Eagle i have a lib where i tested different angles and sizes...
Hope it helps..

Regards, novski

Bourns Encoder.lbr

Most helpful.
I've realised that 16 leds obviously makes it impossible to have  9,12 and 3 o'clock positions.
So either I skip one led  or makes the arch having one extra led on one of the ends.

[Thomasch]

Did you guys know this RGB LED Ring Project?
https://vimeo.com/62498652

[latigid on]

I've chatted with TK about an LED driver chip running SPI, the benefits would be less CPU load and more RGB resolution at the expense of slower transfer rates.

I've ordered a few encoders to test, I'll probably make up a small board with matrix routing and connections to a DOUT module. As always, things get a bit more complicated when you have to route up a control surface PCB with 16 of these on :).

[TK.]

Results: http://midibox.org/forums/topic/19752-rgb-hue-sweep/

 

[latigid on]

Okay, I have the encoders and a 5050 WS2812 LED to test. The problems is that the LED is around 1.65mm tall but the clearance underneath the encoder is only ~1.1mm. 

There are a few ways around this. First, you could solder the encoder in "proud" i.e. raised up above the LED. The neatest way would be to rest the PCB flat against a work surface, insert the encoder and solder from the top. There's a slight risk that your encoder doesn't go in straight. 

Second is to use a washer as a spacer. Nylon is best for non-conductivity. Trouble is there are no washers of the right size. M5 is good on the outside (10mm) but too narrow inside (~5.5mm). M6 is just too small inside, M7 doesn't exist and M8 is far too wide on the outside. We need a washer with 7.5-8mm ID and 10-11mm OD. 2mm thick is best, but 2*1mm would also work. I tried to drill some M5 washers to 7.5mm but the bit chewed up the plastic. I also looked at imperial sizes, 5/16" is fine but 1/2" OD is too much.

An O-ring would at least be thin enough but to get the right height you need 2mm cross section/thickness (stacking them would be difficult). So then we have 7.5mm ID and 11.5mm OD which might just work. I also worry about the lifetime of rubber products.

Lastly one could get pieces of 2mm acrylic laser cut; they could even be square to save on machine time. I wonder about the expense of this though.

So, that's where I'm at at the moment, if you have any ideas please feel free to share.

[TK.]

Using laser cut (e.g. from formulor) was my first idea. There might be some other parts (e.g. LCD/OLED frames) which could be part of the panel as well.

Good idea to use multiple PCBs for the frontpanel.

Best Regards, Thorsten.